Truncation of NH2-terminal Amino Acid Residues Increases Agonistic Potency of Leukotactin-1 on CC Chemokine Receptors 1 and 3*

Leukotactin-1 (Lkn-1) is a human CC chemokine that binds to both CC chemokine receptor 1 (CCR1) and CCR3. Structurally, Lkn-1 is distinct from other human CC chemokines in that it has long amino acid residues preceding the first cysteine at the NH2 terminus, and contains two extra cysteines. NH2-terminal amino acids of Lkn-1 were deleted serially, and the effects of each deletion were investigated. In CCR1-expressing cells, serial deletion up to 20 amino acids (Δ20) did not change the calcium flux-inducing activity significantly. Deletion of 24 amino acids (Δ24), however, increased the agonistic potency ∼100-fold. Deletion of 27 or 28 amino acids also increased the agonistic potency to the same level shown by Δ24. Deletion of 29 amino acids, however, abolished the agonistic activity almost completely showing that at least 3 amino acid residues preceding the first cysteine at the NH2 terminus are essential for the biological activity of Lkn-1. Loss of agonistic activity was due to impaired binding to CCR1. In CCR3-expressing cells, Δ24 was the only form of Lkn-1 mutants that revealed increased agonistic potency. Our results indicate that posttranslational modification is a potential mechanism for the regulation of biological activity of Lkn-1.

Lkn-1 is a human CC chemokine that binds to both CCR1 and CCR3 and induces chemotaxis and calcium influx in human neutrophils, monocytes, eosinophils, and lymphocytes (11). Chemotaxis of neutrophils distinguishes Lkn-1 from other CCR1 agonists such as human macrophage inflammatory protein-1␣ (MIP-1␣) (12). Lkn-1 is a member of a human CC chemokine subfamily that contains four conserved cysteines (11). Lkn-1, however, is distinct from other human CC chemokines in that it has two extra cysteines, which may form a third disulfide bond. Lkn-1 is also distinct from other human CC chemokines in that it has long amino acid residues preceding the first cysteine at the NH 2 terminus. The mature form of Lkn-1 consists of 92 amino acids and has 31 amino acid residues preceding the first cysteine, whereas most human as well as mouse CC chemokines have 10 or fewer amino acid residues preceding the first cysteine (13).
Recombinant Lkn-1 produced in Escherichia coli also shows several distinguished characteristics. In contrast to other CC chemokines such as monocyte chemoattractant protein 1 (MCP-1), recombinant Lkn-1, which contains additional methionine and six histidines at its NH 2 and COOH termini, respectively, shows almost normal biological activities (11). In addition, purified recombinant Lkn-1 undergoes a spontaneous sitespecific cleavage producing a 24-amino acid shorter protein than the intact form of Lkn-1 (11). The biological significance of the spontaneous site-specific cleavage is not known yet. In the present study, we produced a series of NH 2 -terminal deletion mutants of Lkn-1 including the site-specific cleaved form, and the effects of each deletion were investigated in CCR1-or CCR3-expressing cells.
Production and Purification of NH 2 -terminal Deletion Mutants of Lkn-1-A plasmid containing full-length cDNA of Lkn-1 was used as a template in PCR reactions to create Lkn-1 mutants with progressive amino acid deletions at the NH 2 terminus. The sequences of the oligonucleotide primers used for amplification are shown in Table I. All of the forward primers contained overhanging nucleotide sequences for glycine-isoleucine-glutamic acid-glycine-arginine (GIEGR) at the 5Ј end of the target gene, and the reverse primer contained stop codon of Lkn-1 cDNA. The PCR products were cloned into the E. coli expression vector, pET30Xa/LIC (Novagen, Madison, WI), and transformed into E. coli strain BL21(DE3). Expression of Lkn-1 mutant proteins was induced by isopropyl-1-thio-␤-D-galactopyranoside (Sigma). The mutant proteins were purified from bacterial lysate with Ni-NTA spin columns (Qiagen, Chatsworth, CA) according to the manufacturer's instructions. Eluted proteins were folded by gradually removing denaturants in the protein preparation by stepwise dialysis. Overhanging extra amino acid residues at the NH 2 terminus including the polyhistidine tag were removed by cleavage with factor Xa (Novagen). For cleavage of 1 g of target protein, 0.04 unit of factor Xa was added, and the mixture was incubated at 22°C for 7 h. Final purification of the cleaved proteins was performed by chromatographic separation in Superdex Peptide HR 10/30 column (Amersham Biosciences) attached to AKTApurifier (Amersham Biosciences). The proteins were eluted with a 0.02 M phosphate buffer containing 0.25 M NaCl. Finally, purified mutant proteins were free of endotoxin by a Limulus amoebocyte assay (Associates of Cape Cod, Woods Hole, MA). The concentrations of the proteins were determined by micro-bicinchoninic acid assay kit (Pierce) according to the manufacturer's instructions using bovine serum albumin as standard protein.
SDS-PAGE and Western Blot Analysis-SDS-PAGE was performed as described previously (14). The gels were either stained with a 0.025% Coomassie Brilliant Blue R-250 or silver nitrate, or they were transferred onto nitrocellulose membrane (Haake Buchler Instruments, Saddle Brook, NJ). The membrane was probed with polyclonal rabbit anti-Lkn-1 and then with alkaline phosphate-labeled goat anti-rabbit Ig (Bio-Rad). The membrane was developed by the addition of enzyme substrates, 5-bromo-4-chloro-3-indolyl phosphate, and nitro blue tetrazolium (Bio-Rad).
Calcium Influx Assay-Receptor activation was assessed by real time measurement of intracellular calcium concentration in cells labeled with Fura-2/AM (Molecular Probes, Eugene, OR) in MSIII fluorimeter (Photon Technology International, S. Brunswick, NJ) as described previously (4,15). Receptor desensitization was tested by monitoring intracellular calcium changes in cells upon repeated chemokine stimulation at 100-s intervals. Results were expressed as excitation ratios at 340 and 380 nm.
Chemotaxis Assay-Chemotactic activities were performed in a 48microwell Boyden chamber (Neuroprobe, Cabin John, MD) as described previously (15). The lower wells were filled with 27 l of buffer alone or SCHEME 1. Schematic diagram of the NH 2 -terminal deletion mutants of Lkn-1 and the expression vector. A, DNAs for the NH 2terminal deletion mutants were generated by polymerase chain reactions with primers (shown in Table I) from full-length cDNA of Lkn-1. B, the PCR products were cloned into the E. coli expression vector pET30Xa/LIC, which encodes fusion protein that has the 6-amino acid His-tagged and S-tagged sequences at the upstream of the target protein. 5Ј-GGTATTGAGGGTCGCCAGTTCACAAATGATGCAGAG-3Ј

5Ј-AGAGGAGAGTTAGAGCCTTATATTGAGTAGGGCTTCAGC-3Ј
a The factor Xa recognition site is underlined.
with buffer containing chemokines, and the upper wells were filled with 50 l of cell suspension (6ϫ10 5 cells/ml). The two wells were separated by a fibronectin (Sigma)-coated polyvinylpyrrolidone-free polycarbonate filter (Neuroprobe) with 10-m pores. Human MIP-1␣ and human eotaxin were purchased from PeproTech Inc. (Rocky Hill, NJ).
Receptor Binding Assay-Lkn-1 mutant proteins were labeled radioactively with 125 I using the chloramine T method (16). The specific activities of the labeled Lkn-1 mutants were ϳ 8ϫ10 7 cpm/g protein. CCR1-or CCR3-transfected cells were suspended in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum at a concentration of 2ϫ10 7 cells/ml; 100 l of the suspension was added to each tube. Saturation binding assays were performed in a total volume of 0.2 ml containing 2ϫ10 6 cells and 0.04 to ϳ10 nM 125 I-labeled Lkn-1 mutants. Competition binding experiments were performed under the same conditions using 1 nM 125 I-labeled ⌬0 and variable concentrations of competitors. The nonspecific binding was estimated by measuring the binding of the labeled ligand in the presence of 100-fold excess of unlabeled ligand. Samples were incubated at 4°C for 90 min with continuous rotation. Incubation was terminated by centrifuging the cell suspension over 1 ml of 10% sucrose (Sigma) cushion. Cell pellets were cut from the tubes, and cpm were counted using a gamma counter.

Production and Purification of NH 2 -terminal Deletion Mutants-
The intact form of Lkn-1 and its mutants lacking NH 2terminal 5, 10, 15, 20, 24, 27, 28, and 29 amino acid residues, respectively, were produced in E. coli using the T7 polymerasebased expression vector, pET-30 Xa/LIC. Scheme I depicts the NH 2 -terminal deletion mutants of Lkn-1 produced as well as the expression system used in the present study.
Recombinant fusion proteins were readily detected in bacterial lysate (Fig. 1A). The fusion proteins were purified from the bacterial lysate by nickel-chelating resin (Fig. 1B). The purified fusion proteins were strongly reactive with polyclonal rabbit anti-Lkn-1 antibody (Fig. 1C).
Purified fusion proteins were then cleaved with the protease, Factor Xa, which recognizes GIEGR sequences linked directly to the NH 2 terminus of the target protein. Factor Xa cleaved most of the fusion proteins almost completely. Factor Xa, however, was less efficient in cleaving fusion proteins for ⌬27, ⌬28, and ⌬29. Thus, final purification and removal of uncleaved fusion proteins were performed by a chromatographic separation. Fig. 2A shows a representative chromatogram of the ⌬28 fusion protein that was digested with Factor Xa. The protein in the single dominant peak in Fig. 2A was collected and found to be a homogeneous protein that was cleaved with Factor Xa (Fig. 2B, lane 2). Likewise, other fusion proteins were also cleaved with Factor Xa, and then the cleaved proteins were purified by a chromatographic separation. The purified proteins contained undetectable level of endotoxin (data not shown).
Induction of Ca 2ϩ Mobilization in CCR1-and CCR3-expressing Cells-Recombinant proteins of the NH 2 -terminal deletion mutants were compared for calcium mobilization in CCR1 transfectant cells. When calcium flux-inducing activity was compared at 100 nM concentration, which is the concentration shown to induce maximal calcium flux by ⌬0, it was evident that agonistic potency decreases dramatically between ⌬28 and ⌬29 (Fig. 3A). To further confirm that deletion of one more amino acid from ⌬28 results in an almost complete loss of agonistic potency, calcium flux-inducing activity of ⌬29 was compared at several different concentrations with that of ⌬28 as well as that of MIP-1␣. As shown in Fig. 3B, calcium fluxinducing activity of ⌬28 was dose-dependent, reaching a plateau at 10 nM. In contrast, ⌬29 elicited maximal response at Recombinant fusion proteins were cleaved with Factor Xa and then separated by gel filtration using a Superdex peptide HR 10/30 column (Amersham Biosciences) to remove tagged peptides and uncleaved fusion proteins. Here, we show purification of ⌬28 fusion proteins as a representative example. A, Factor Xa-treated ⌬28 fusion proteins were separated by gel filtration on a Superdex peptide HR 10/30 column, and the eluted proteins were detected at ϭ 214 nm. B, the protein in the single dominant peak in A was collected and separated in SDS-PAGE, and the gel was stained with silver nitrate. Lane 1 shows total proteins before the chromatographc separation, and lane 2 shows the finally purified Factor Xa-treated proteins. Std indicates the protein size marker. 100 nM, but the level was at most comparable with that of 0.1 nM of ⌬28. These results show that at least 3 amino acid residues preceding the first cysteine are required for the agonistic activity.
Potential differences in the agonistic potency were further examined for all of the NH 2 -terminal deletion mutants by varying protein concentration. We found that deletion of NH 2 -terminal amino acid residues up to 20 did not change agonistic potency for CCR1; the calcium flux-inducing activities of ⌬5, ⌬10, ⌬15, and ⌬20 were almost the same as that of ⌬0 (Fig. 3,  C and D). Lkn-1 lacking NH 2 -terminal 24 amino acids (⌬24), however, induced robust calcium flux responses compared with ⌬0. ⌬24 induced maximal calcium flux response at ϳ1 nM, which is an ϳ100-fold lower concentration than that required to induce similar level of response for ⌬0 (Fig. 3D). EC 50 of ⌬24 and ⌬0 was ϳ0.2 and ϳ4.0 nM, respectively. Deletion of 3 or 4 more amino acid residues from ⌬24 (i.e. ⌬27 and ⌬28) also increased agonistic potency to the same level shown by ⌬24. The calcium flux-inducing activity of ⌬28 as well as ⌬24 appears to be stronger than that of MIP-1␣ (Fig. 3B).
Agonistic potency on CCR3 was also examined for all of the NH 2 -terminal deletion mutants using CCR3 transfectant cells, and some of the data are shown in Fig. 4. In CCR3 transfectant cells, ⌬24 was the only form of the NH 2 -terminal deletion mutants that showed enhanced agonistic potency; the EC 50 of all the other NH 2 -terminal deletion mutants as well as ⌬0 was in the range of 40 to ϳ55 nM. In contrast, the EC 50 of ⌬24 was ϳ22 nM. The agonistic potency of ⌬24, however, was lower than that of eotaxin, which showed a peak response at 50 nM with an EC 50 of ϳ18 nM. At a 50 nM concentration, ⌬24 exhibited ϳ50% of the calcium flux response to eotaxin. It is noteworthy that the EC 50 of ⌬24 on CCR1 was ϳ 0.2 nM, which is ϳ100-fold lower than that on CCR3.
Desensitization Experiments-To test whether loss of agonistic activity of ⌬29 is due to inability to bind to CCR1, desensitization experiments were performed in CCR1 transfectant cells. As shown in the upper panel of Fig. 5A, MIP-1␣ was not able to desensitize CCR1 transfectant cells completely to subsequent exposure to ⌬0 or ⌬28. However, stimulation of CCR1 transfectant cells with ⌬0 or ⌬28 completely abolished responsiveness to a second stimulation with MIP-1␣. These data indicate that ⌬28, as well as ⌬0, shares receptors with MIP-1␣ and has a stronger desensitizing capability than MIP-1␣. Then, we examined whether ⌬29 could desensitize ⌬28. Stimulation of CCR1 transfectant cells with ⌬29 did not affect the ability of ⌬28 to induce the subsequent calcium response, even at a concentration as high as 400 nM (Fig. 5A). These results suggest that ⌬29 may not be able to bind to CCR1.
Desensitization experiments were also performed in CCR3 transfectant cells to compare desensitizing capability of ⌬24 with that of eotaxin. As shown in Fig. 5B, stimulation of CCR3 transfectant cells with eotaxin abolished responsiveness to a subsequent stimulation with ⌬24. In contrast, stimulation with ⌬24 did not completely abolish the ability of eotaxin to induce a calcium response.
Induction of Chemotaxis in CCR1-and CCR3-expressing Cells-Chemotactic activities of the NH 2 -terminal deletion mutants of Lkn-1 were evaluated on CCR1 transfectant cells, and the results are shown in Fig. 6A. Consistent with the calcium flux results, chemotactic activity decreased dramatically when 29 amino acids were deleted. ⌬29 did not show significant chemotactic activity even at a concentration of 10 nM.
In CCR3 transfectant cells, ⌬24 was again the only form of the NH 2 -terminal deletion mutant of Lkn-1 that showed increased agonistic activity (Fig. 6B). The chemotactic activity of ⌬24, however, was lower than that of eotaxin. The EC 50 of ⌬24 was 33.0 Ϯ 2.5, whereas that of eotaxin was 13.0 Ϯ 1.2. Furthermore, at a concentration of 100 nM, at which eotaxin showed a peak response, ⌬24 exhibited 58% of the chemotactic activity of eotaxin.
Receptor Binding Assay-Binding affinities of Lkn-1 mutant proteins with CCR1 and CCR3 were determined with 125 Ilabeled ligands (Fig. 7). In this experiment, we focused mainly on three proteins: ⌬0, which is the intact form of Lkn-1; ⌬24, which is the deletion mutants with enhanced agonistic activities on CCR3 as well as CCR1; and ⌬29, which essentially lacks the biological activity. Fig. 7, A and B, depicts the binding isotherm. Conversion of the data by Scatchard analysis revealed a K d of 1.00 Ϯ 0.02 nM (⌬0) and 0.38 Ϯ 0.01 nM (⌬24) to the CCR1 (Fig. 7A), and 2.41 Ϯ 0.08 nM (⌬0) and 1.05 Ϯ 0.09 nM (⌬24) to the CCR3 (Fig. 7B). The relative affinity of the Lkn-1 mutants was investigated further in a cross-competition binding assay using CCR1 transfectant cells (Fig. 7C). ⌬24 was more effective in inhibiting the binding of 125 I-labeled ⌬0 with CCR1 than ⌬0. The IC 50 of ⌬24 was ϳ26 nM, whereas that of ⌬0 was ϳ42 nM. ⌬29, at the highest concentration tested, inhibited the binding of 125 I-labeled ⌬0 by only 20%. DISCUSSION It has been found that recombinant Lkn-1 produced in insect cells undergoes a spontaneous site-specific cleavage at the NH 2 FIG. 6. Chemotactic activity of the NH 2 -terminal deletion mutants. Chemotaxis assays were performed using a Boyden chamber. A, chemotactic activity of the deletion mutants was compared in a doseresponse study using CCR1-HOS cells, and some of the representative data are shown. MIP-1␣ was served as internal standards. B, chemotactic activity of ⌬0, ⌬10, ⌬24, and ⌬28 was compared in a dose-response study with that of eotaxin. The results show the mean Ϯ S.D. of three separate experiments. terminus, producing a 24-amino acid shorter protein than the intact form (11). The significance of this autolysis, however, has remained unknown. In the present study, the role of the NH 2terminal domain for agonistic activity was studied with a series of NH 2 -terminal deletion mutants of Lkn-1, including the spontaneously cleaved form (⌬24). Our preparations of NH 2 -terminal deletion mutants of Lkn-1 do not have any foreign amino acid residues at either the NH 2 or COOH terminus of the recombinant proteins. Although recombinant Lkn-1, which has additional methionine and six histidines at its NH 2 terminus, has biological activities (11), we found that truncation of the foreign amino acid residues at the NH 2 terminus of the recombinant proteins increases the biological activity by ϳ10-fold (data not shown).
In contrast to other CC chemokines such as MCP-1 and RANTES, deletion of NH 2 -terminal amino acids up to 20 residues from the natural form of Lkn-1 did not cause noticeable alterations in agonistic potency on CCR1. This feature is unique for Lkn-1, because for CC chemokines such as MCP-1 and RANTES, minimal truncation or modification of the first few NH 2 -terminal amino acids leads to significant changes in receptor binding and functional activity (17)(18)(19)(20). Deletion of the pyroglutamate residue at the NH 2 terminus of the natural form of MCP-1 results in an at least 50-fold decreases in agonistic activity on monocytes and basophils (17,18). Deletion of 2 amino acid residues, MCP-1-(3-76), leads to total loss of agonistic activity on eosinophils and basophils (18). RANTES loses agonistic potency and becomes a potent antagonist of chemokine binding when the first amino acid residue has been modified artificially by the addition of methionine or treatment with aminooxypentane (19,21). Deletion of NH 2 -terminal amino acid residues can even result in changes in receptor specificity (18,22). In particular, MCP-1 acquires agonistic activity on CCR3 only when the first residue at the NH 2 terminus, pyroglutamate, is deleted (18). Similar to artificially modified chemokines, posttranslationally modified natural forms of MCP-1 such as MCP-1-(5-76) and MCP-1-(6 -76) are also devoid of bioactivity (20). The naturally cleaved form of MCP-2, MCP-2-(6 -76), is also devoid of bioactivity and blocks the chemotactic effects of MCP-2 as well as that of MCP-1, MCP-3, and RANTES (20). Thus, the integrity of the NH 2terminal region of CC chemokines appears to be critical for receptor binding and biological function. In fact, the search for NH 2 -terminal variants as receptor antagonists has been one of the major areas of interest since the discovery that chemokines inhibit HIV-1 infection (9,(23)(24)(25).
The situation is quite different for Lkn-1. Our results show that deletion of up to 20 amino acids from the intact form of Lkn-1 does not affect agonistic potency on CCR1. More interestingly, recombinant Lkn-1, which contains additional methionine and six histidines at its NH 2 terminus, is only 10-fold less active the than intact form of Lkn-1. These observations suggest that the receptor binding and biological function of Lkn-1 is more dependent on the downstream amino acid residues of the NH 2 -terminal domain. Deletion of more amino acid residues, 24 amino acids (⌬24), increases the calcium flux-inducing activity almost 100-fold on CCR1 transfectants. These results may explain why recombinant Lkn-1 undergoes a spontaneous site-specific cleavage at the NH 2 terminus producing a 24amino acid shorter protein than the intact form (10). For Lkn-1, posttranscriptional modification may be a mechanism to augment chemokine potency on CCR1. This is also true for CCR3; ⌬24 shows increased calcium flux-inducing activity in CCR3 transfectants compared with that of the intact form of Lkn-1. Deletion of 27 or 28 amino acids (⌬27 or ⌬28) also produces a protein that has increased calcium flux-inducing activity on CCR1 transfectants. Deletion of 29 amino acid residues (⌬29), however, abolishes the calcium flux-inducing activity almost completely, showing that at least 3 amino acid residues preceding the first cysteine are essential for the biological activity of Lkn-1.
Because deletion or proteolytic cleavage of the NH 2 -terminal region usually results in derivatives that still recognize the receptor but do not induce functional responses, we were curious to know whether ⌬29 acts as an antagonist of Lkn-1. This issue was addressed by receptor binding experiments as well as desensitization experiments. As shown in Fig. 6A, ⌬29 was unable to desensitize calcium flux-inducing activity of ⌬28 even at 400 nM concentration. Furthermore, receptor binding experiments showed that ⌬29 was not able to bind to CCR1 effectively. Thus, it was obvious that deletion of 29 amino acids at the NH 2 terminus inactivates the receptor binding capability of Lkn-1. This feature is unique for Lkn-1 in that it does not produce antagonists by truncations of the NH 2 -terminal domain.
Although cleavage of the NH 2 -terminal region increases the agonistic potency on CCR1, the length of the NH 2 -terminal region does not proportionally affect agonistic potency; ⌬24, ⌬27, and ⌬28 show similarly potent biological activity in both calcium flux assays and migration assays. This feature is in contrast with that of hemofiltrate CC chemokine (HCC)-1, a recently cloned CC chemokine that is structurally similar to MIP-1␣ (26). Comparison of the three NH 2 -terminal truncated variants of HCC-1 revealed that the rank order of potency was inversely correlated with the length of the protein. Thus, the shortest form of HCC-1 was the most potent, and the longest form of HCC-1 was the least potent (27).
To examine whether increased agonistic potency shown by the deletion mutants is due to increased binding affinity to receptors, we also performed receptor binding assays for ⌬0 and ⌬24. Based on the K d values, ⌬24 appeared to have an ϳ2.6-fold higher binding affinity on CCR1 than ⌬0. Crosscompetition experiments using labeled ⌬0 also showed that ⌬24 had higher binding affinity on CCR1 than ⌬0. Because both ⌬27 and ⌬28 have similarly strong agonistic potency as ⌬24 on CCR1, it is reasonable to assume that ⌬27 and ⌬28 would also have similar binding affinity as ⌬24 on CCR1. In CCR3 transfectants, ⌬24 showed an ϳ2.3-fold higher binding affinity than ⌬0. Thus, it appears that the increased agonistic potency of the NH 2 -terminal deletion mutants is due, at least in part, to the increased binding affinity to receptors.